Estimation Of Parameters In High Speed Vehicle Based On Multi-filter

High-speed railway has been developing vigorously in China and the related technologies are constantly emerging. At present, the level of China’s vehicle manufacturing has been at the forefront of the world. The vehicle becomes the main exported and welcomed product all around the world. However, the fault detection method of high speed railway vehicle is still relatively primitive, which needs further improvement.To improve the fault detecting techniques, a method based on parameter estimation is introduced. By conducting the topological analysis, a simplified model is obtained and the dynamic analysis is carried out to get its mathematical model to estimate the parameters. However, big differences are showed between the simplified model and the practical model, so the anti-hunting damper, the air spring and other parts of the train are modeled in detail. Thus, a switching model of CRH2 Multiple Units is built. The simulation of the model shows a better result.The simplified model of the train is a linear model, yet it turned into a nonlinear model after parameter estimation. According to the model of Gaussian white noise, both extended Kalman filter and Rao-Blackwellised particle filter show significance. However, the effect of parameter estimation is not satisfactory, especially when the excitation is changed to the actual track spectrum. Due to this reason, the standard particle filter is introduced to solve the problem. To deal with the switching model, the multi-filter joint strategy is proposed to estimate the parameters of each subsystem of the switching model and exchange data. It is concluded that the strategy suggested better results on parameters estimation of the model.The fault detection method based on parameter estimation in this paper is quite new in the field of train fault detection, which is considered to have a very wide application prospect.